Multi-objective reactive power planning based on fuzzy clustering and learning automata

Reactive power planning (VAR Planning) is one of the most challenging issues in the domain of power system research. It is a mixed integer nonlinear optimization problem with a large number of variables and uncertain parameters. In this paper, first, the fuzzy clustering method is employed to select candidate locations for installing new shunt VAR sources. Specifically, U/U
₀
index, G index, and a critical voltage magnitude index are employed to form data matrix of fuzzy clustering. Second, a multi-objective optimization model is proposed for VAR optimization considering generation cost, VAR device cost, voltage stability and active power loss. A P-model learning automata algorithm is used to provide the multi-objective optimization solutions. Test results on a IEEE 57-bus system clearly demonstrate that a learning automata is a feasible method to produce a multi-objective trade-off analysis; and the combination of fuzzy clustering and learning automata can be a prospective method for multi-objective reactive power planning.